large object regression tests, take two

Started by Jeremy Drakealmost 20 years ago6 messagespatches
Jump to latest
#1Jeremy Drake
pgsql@jdrake.com

This is the latest version of the large object regression test I have been
working on. Note that a prerequisite for this version of the test is the
patch I made to psql to make it not output on \lo_* commands in quiet mode
is required (also attached, it's small).

Sorry that it still makes use of the sed trickery like the copy test does,
but:

1) I think the server side lo_(import|export) functions need to be tested
as well as the psql variants

2) ISTM that making assumptions about the working directory of psql
during
the regression tests would open a can of worms, especially wrt VPATH
builds where the data files could be in a completely separate tree from
the regression tests.

Thoughts?

--
Why did the Roman Empire collapse?
What is the Latin for office automation?

Attachments:

psql-largeobj-output.patchtext/plain; charset=US-ASCII; name=psql-largeobj-output.patchDownload+44-7
largeobj-regress-v2.patch.gzapplication/octet-stream; name=largeobj-regress-v2.patch.gzDownload
#2Bruce Momjian
bruce@momjian.us
In reply to: Jeremy Drake (#1)
Re: large object regression tests, take two

This has been saved for the 8.3 release:

http://momjian.postgresql.org/cgi-bin/pgpatches_hold

---------------------------------------------------------------------------

Jeremy Drake wrote:

This is the latest version of the large object regression test I have been
working on. Note that a prerequisite for this version of the test is the
patch I made to psql to make it not output on \lo_* commands in quiet mode
is required (also attached, it's small).

Sorry that it still makes use of the sed trickery like the copy test does,
but:

1) I think the server side lo_(import|export) functions need to be tested
as well as the psql variants

2) ISTM that making assumptions about the working directory of psql
during
the regression tests would open a can of worms, especially wrt VPATH
builds where the data files could be in a completely separate tree from
the regression tests.

Thoughts?

--
Why did the Roman Empire collapse?
What is the Latin for office automation?

Content-Description:

[ Attachment, skipping... ]

Content-Description:

[ Attachment, skipping... ]

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

--
Bruce Momjian bruce@momjian.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

#3Bruce Momjian
bruce@momjian.us
In reply to: Jeremy Drake (#1)
Re: large object regression tests, take two

Is this patch ready for application?

---------------------------------------------------------------------------

Jeremy Drake wrote:

This is the latest version of the large object regression test I have been
working on. Note that a prerequisite for this version of the test is the
patch I made to psql to make it not output on \lo_* commands in quiet mode
is required (also attached, it's small).

Sorry that it still makes use of the sed trickery like the copy test does,
but:

1) I think the server side lo_(import|export) functions need to be tested
as well as the psql variants

2) ISTM that making assumptions about the working directory of psql
during
the regression tests would open a can of worms, especially wrt VPATH
builds where the data files could be in a completely separate tree from
the regression tests.

Thoughts?

--
Why did the Roman Empire collapse?
What is the Latin for office automation?

Content-Description:

[ Attachment, skipping... ]

Content-Description:

[ Attachment, skipping... ]

---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

#4Alvaro Herrera
alvherre@2ndquadrant.com
In reply to: Bruce Momjian (#3)
Re: large object regression tests, take two

Bruce Momjian wrote:

Is this patch ready for application?

I think the sed usage would cause problems for the VC++ builds. It
would be good if the files could be reformulated as files in the input,
output or data directories; pg_regress would take care to generate the
files as needed.

---------------------------------------------------------------------------

Jeremy Drake wrote:

This is the latest version of the large object regression test I have been
working on. Note that a prerequisite for this version of the test is the
patch I made to psql to make it not output on \lo_* commands in quiet mode
is required (also attached, it's small).

Sorry that it still makes use of the sed trickery like the copy test does,
but:

1) I think the server side lo_(import|export) functions need to be tested
as well as the psql variants

2) ISTM that making assumptions about the working directory of psql
during
the regression tests would open a can of worms, especially wrt VPATH
builds where the data files could be in a completely separate tree from
the regression tests.

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

#5Jeremy Drake
pgsql@jdrake.com
In reply to: Alvaro Herrera (#4)
Re: large object regression tests, take two

On Thu, 8 Feb 2007, Alvaro Herrera wrote:

Bruce Momjian wrote:

Is this patch ready for application?

It already has been.

I think the sed usage would cause problems for the VC++ builds. It
would be good if the files could be reformulated as files in the input,
output or data directories; pg_regress would take care to generate the
files as needed.

It Just Worked with the changes made to pg-regress to support the other,
similar tests (ie, copy).

---------------------------------------------------------------------------

Jeremy Drake wrote:

This is the latest version of the large object regression test I have been
working on. Note that a prerequisite for this version of the test is the
patch I made to psql to make it not output on \lo_* commands in quiet mode
is required (also attached, it's small).

Sorry that it still makes use of the sed trickery like the copy test does,
but:

1) I think the server side lo_(import|export) functions need to be tested
as well as the psql variants

2) ISTM that making assumptions about the working directory of psql
during
the regression tests would open a can of worms, especially wrt VPATH
builds where the data files could be in a completely separate tree from
the regression tests.

--
The human mind treats a new idea the way the body treats a strange
protein -- it rejects it.
-- P. Medawar

#6Bruce Momjian
bruce@momjian.us
In reply to: Jeremy Drake (#5)
Re: large object regression tests, take two

Jeremy Drake wrote:

On Thu, 8 Feb 2007, Alvaro Herrera wrote:

Bruce Momjian wrote:

Is this patch ready for application?

It already has been.

Ah, I see it now. Strange I missed seeing that. Thanks.

---------------------------------------------------------------------------

I think the sed usage would cause problems for the VC++ builds. It
would be good if the files could be reformulated as files in the input,
output or data directories; pg_regress would take care to generate the
files as needed.

It Just Worked with the changes made to pg-regress to support the other,
similar tests (ie, copy).

---------------------------------------------------------------------------

Jeremy Drake wrote:

This is the latest version of the large object regression test I have been
working on. Note that a prerequisite for this version of the test is the
patch I made to psql to make it not output on \lo_* commands in quiet mode
is required (also attached, it's small).

Sorry that it still makes use of the sed trickery like the copy test does,
but:

1) I think the server side lo_(import|export) functions need to be tested
as well as the psql variants

2) ISTM that making assumptions about the working directory of psql
during
the regression tests would open a can of worms, especially wrt VPATH
builds where the data files could be in a completely separate tree from
the regression tests.

--
The human mind treats a new idea the way the body treats a strange
protein -- it rejects it.
-- P. Medawar

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +